1. Manitoba doesn't need another gas plant. It needs heat that works twice.
Existing Brandon units were failed or limited during 3 of the 10 highest multi-day winter peak events examined.
Providing quick heating during harsh winter mornings is essential. According to CBC News PD, Manitoba Hydro told the Public Utilities Board that it reduced the output of Brandon’s gas turbines during three of the ten biggest multi-day winter peak events in the last decade due to supply chain issues.
Here is what the cold-wave record reveals, and what it does notDespite this track record, Hydro is requesting about $3 billion for new turbines at Brandon. New machines will perform better than the existing units, but that ask still has to show that they will start and run on the mornings that matter. Fuel supply and cold-weather starts still need specific modelling.
Manitoba already has a solution that works for several days: its reservoirs hold a large amount of stored energy, even under drought planning conditions PD A. The 2030 problem is not a shortage of energy across a winter. It is a shortage of power in a handful of hours, and the expected shortfall of 600 MW in a drought year PD is real.
That distinction is what makes this peak worth attacking from both ends. A gas plant is sized to whatever demand turns up on the coldest morning, which means the cheapest megawatt available is the one that never arrives. Keeping the gas furnace on the last few degrees of a −35 °C night is how it never arrives: the heat pump still carries most of the season, while the hour that decides the size of the plant stays on the gas system. That is Hydro’s own dual-fuel assumption, and on the homes it covers it adds about 0 MW at the peak. Pull those furnaces out without a replacement plan and the same homes can add hundreds of megawatts in the one hour the system can least afford it. Building 750 MW of turbines to serve a peak that heating policy helped create is solving the problem backwards.
So the first job is to stop adding to that hour, and the second is to shave what is left of it. We believe a hybrid wastewater heat recovery model can act as a 2030 coincident-power bridge for that hour, and that gas is not needed if Order 1 in our asks shows the hour without new turbines. The key question is not whether to act, but where to invest.
Data sources — registry entries this figure is built from: peak_record_mw, peak_record_temp_c, winter_design_temp_c, shortfall_2030_mw, hub_net_electric_mwCurve shape from winter_load_base_mw, winter_heat_threshold_c, winter_heat_beta_mw_per_c.
2. Two $3-billion choices, one wastes heat forever
Both decisions are still open. Brandon is before the Board, and the North End rebuild has not yet closed the chance to recover its heat.
The province is currently debating a $3-billion gas plant in Brandon. Winnipeg is already building a $3.17 billion project: modernising the North End Water Pollution Control Centre—the largest capital project in the city's history. NEWPCC already treats about 70% of Winnipeg's wastewater. Sewage carries low-grade heat every hour of every season. Vancouver's False Creek has heated buildings this way for more than a decade. On a −35 °C morning, every megawatt of that heat that reaches a radiator is one less megawatt Hydro must generate for electric heating. Use the same energy, twice.
If we miss the window then Winnipeg gets a plant that treats water well and wastes heat forever. But if we act now while contracts are open, then the North End can host low-carbon heat for a generation—without asking ratepayers to fund another fossil fuel plant — and we can start saving and better accounting for our energy.
| Decision | Brandon gas | NEWPCC clean hybrid |
|---|---|---|
| Capital cost | $3.0B | $1.95B gross ($1.50B if ITC confirmed) |
| Overnight $/kW | $4,000/kW nameplate | $2,834/kW gross / $2,180/kW net at 688 MW planning |
| Winter resource stack | 750 MW nameplate; 525 MW in a 30% derate scenario only | 413 / 688 / 788 MW electric — conservative / planning / full |
| District heat | None | 50 MWth thermal effluent stage (+95 MWth compute, not winter-firm until metered); +38 MW net electric credit |
| 20-year total | $4.5-6.9B | $1.5-1.8B net of ITC |
Under our modelled cost bands, the difference is $2.7–5.4B over twenty years E A
Data sources — registry entries this figure is built from: gas_capex_b, hybrid_capex_gross_b, hybrid_capex_net_b, gas_20yr_low_b, gas_20yr_high_b, hybrid_20yr_low_b, hybrid_20yr_high_b · computed by system_b_ledger.py
That is our solution, a heat-pump and battery answer to a heat problem, not a gas-plant answer to a spreadsheet problem.
3. Our plan: sewage heat, batteries, demand response, interties
Then Hydro can put Conawapa-class hydro back on the capacity path instead of a peaker.
Manitoba already stores the energy it needs in its reservoirs. The waste happens at the point of use, where an electron is spent once on one job and its work ends there. In our hybrid “heat hub” model, the same electron works twice. Sewage-source heat pumps on warm treated effluent run at a heat-pump COP of 4.13 (49.6 MWth delivered ÷ 12 MW compressor electricity). That is a different machine from an air-source heat pump on a −35 °C morning, where COP falls toward resistance backup; the hub is built for the cold hour, not against it. That is our “heat machine” — 12 MW of compressor electricity. Around it, the planning stack is batteries, prospective demand response, and intertie (imports).
On a -35°C morning in Winnipeg, when the grid is stretched to its breaking point, recovered sewage heat keeps warming buildings that would otherwise draw electric resistance heat. This isn't a single giant machine replacing Brandon's gas station—it's four modular pieces on the winter stack: sewage heat recovery, batteries, prospective demand response, and interties. Public compute is a third book that can add heat and revenue, but it's kept off that stack.
Hydro’s current planning estimate for Brandon is about $3 billion for a nameplate capacity of 750 MW, but this is not the final construction cost, and past events at Brandon suggest that tail-hour availability should be directly modelled, though those events are not themselves an ELCC calculation. Public Utilities Board materials show that the estimate for cost has also changed. This plant also does not recover heat.
Our NEWPCC hybrid hub, on the other hand, uses modular components that can be added over time. Its winter resource stack ranges from 413 to 788 MW, depending on intertie supply, and that is without a compute credit. The hub also provides 49.6 MWth of district heating from effluent (rounded to about 50 MWth in planning), which helps lower peak demand for electric heating. Public compute is tracked separately as a third book as it can make the loop more useful.
Think of five pieces working together, not one giant machine replacing Brandon.
Data sources — registry entries this figure is built from: hub_thermal_mwth, hub_net_electric_mw, scenario_reference_mw
-
Sewage heat at NEWPCC (the base).
Pump heat already in the North End plant’s treated water into a district loop: the 38 MW grid benefit is an upper bound — it counts only if the 50 MWth displaces coincident electric resistance, and buildings still on gas would add the 12 MW compressor load instead — and the water stays warm in January with or without compute.
-
300 MW batteries (4 hours).
A 300 MW, four-hour pack covers 7–9 a.m. and 5–7 p.m. — 300 MW × 2 h + 300 MW × 2 h = 1,200 MWh — shaving the two daily spikes so hydro is not recalled to heat houses that still have furnaces, and holding power for the hours MISO will pay. The pack must recharge between critical days, Hydro has to show it can, and an 85% haircut is not dependable capacity.
-
150 MW prospective demand response.
Pay willing customers to drop the worst hours: 150 MW is a planning target E for residual peak after the loop, not a contracted resource, and not the 38 MW already displaced.
-
Interties up to 300 MW.
Use the high-voltage lines to Saskatchewan, Ontario, and MISO: planning counts 200 of 300 MW as 300 × 2/3 — a cold-hour convention, not an ELCC study; 29% of 688, 488 MW without it, zero until Order 1 proves a neighbour can deliver. That 200 MW of dependable imports during Manitoba’s winter peak has not been demonstrated in this filing; intervenor evidence in the current proceeding may be stronger, it is not a PUB ruling, imports must not be counted twice against Hydro’s baseline, and reservoir or export benefits from lower domestic demand cannot be counted in the same hour as an import.
-
Small public compute, a potential benefit off the winter stack.
Small 10–20 kW public nodes could sit on the heat loop, heat-first, off at 7–9 a.m. and 5–7 p.m., metered in public: a trial would prove those rules and could add heat and revenue, still 0 MW on 413 / 688 / 788. If it grew, the same loop could take 95 MWth of reject heat, still 0 MW winter-firm until metered — compute does not eat the 600 MW shortfall.
A multi-part public hybrid gives Manitoba more energy choices than a single gas plant. Batteries, sewage-heat recovery, demand response, and interties can be sized and sequenced; none of them requires pouring one irreversible bet before the arithmetic is tested.
One hub, three books
Heat, electricity, and compute are not the same thing, so they do not share one total.
The North End Water Pollution Control Centre has warm treated effluent year-round. We want to take heat from that water, and add it to a better-managed greener portfolio.
Heat is counted in MWth. The electricity used to move it is counted in MW. Those are different units, so we don’t add them as if they were the same thing — and that’s a part of our solution.
To calculate how much heat we have to work with and to plan the best solution stack — we need to work out the math on five variables. Our first step is to calculate how much heat is in the water. NEWPCC treats about 195 ML/d. Heat pulled from a stream is its mass flow times specific heat times the temperature drop: so the more you cool a liquid, the more heat you get. For this plant’s particular flow, cooling by 1 °C yields about 9.45 MWth. The ledger rounds that down to 9.4 E:
1 °C of cooling = 9.4 MWth
We do not extract all of the heat out of the treated water, since it returns to the river. River-safe practice is a 4 °C drop. That is the environmental limit this proposal uses. Four degrees, not five, not ten.
4 × 9.4 = 37.6 MWth E
The second number we need is the electricity to run the compressors, which is 12 MW.
Third, the heat delivered to the district loop. A heat pump does not create energy. It moves the 37.6 MWth already in the water, and the compressor work shows up as heat too, so that number is:
37.6 + 12 = 49.6 MWth
Planning rounds 49.6 to 50 MWth.
Fourth, the coefficient of performance. COP is delivered heat divided by compressor electricity:
49.6 ÷ 12 = 4.13
Heat-pump COP: 4.13
This means that for every megawatt of electricity the compressors draw, the loop delivers 4.13 megawatts of heat. No new energy was created. The heat pump transferred existing heat.
The fifth number tells us what that COP does to the electric peak. The electrical benefit is reduced, because we still have to buy the 12 MW, so the net electric reduction is:
50 − 12 = 38 MW
This equation treats the 50 MWth as displacing resistance coils on the peak hour. Coincidence is 100%. That 38 MW is an upper bound. It occurs only if all 50 MWth replaces coincident electric-resistance heating. If the connected buildings currently use natural gas, the project instead adds roughly 12 MW of electric load while reducing gas consumption. Sensitivity: 80% coincidence gives 28 MW; 60% coincidence gives 18 MW.
The 38 MW is the only instance where recycled heat is included in the electric stack.
From there, the planning stack is:
38 + 300 + 150 + 200 = 688 MW
The 688 MW planning case requires 200 MW of intertie availability. Without that import credit, the stack is:
688 − 200 = 488 MW
The 300 MW is a planning ceiling for import on the existing Saskatchewan, Ontario, and MISO ties — an assumption, not a sum of named line ratings. The planning case takes two-thirds of that ceiling. That haircut is an assumption to be validated, not an ELCC result:
300 × 2/3 = 200 MW E
As a share of the planning stack:
200 ÷ 688 ≈ 29%
Deliverability during peak cold hours must be demonstrated (Order 1); it is not recorded as firm capacity.
688 MW is the planning case we just stacked. The other two cases turn two knobs: demand response, and how much of that 300 MW intertie we count.
Conservative counts no intertie, and only half the demand response:
38 + 300 + 75 + 0 = 413 MW
If the hub is also offline, drop the 38:
300 + 75 = 375 MW
Full counts the whole 300 MW intertie instead of two-thirds:
38 + 300 + 150 + 300 = 788 MW
The three cases therefore look like this:
- 413 MW — conservative: 38 MW heat, 300 MW battery, 75 MW demand response, and 0 MW intertie. If the hub is offline, the conservative case is 375 MW (batteries and half demand response only).
- 688 MW — planning: 38 MW heat, 300 MW battery, 150 MW demand response, and 200 MW intertie.
- 788 MW — full: 38 MW heat, 300 MW battery, 150 MW demand response, and 300 MW intertie.
Data sources — registry entries this figure is built from: scenario_conservative_mw, scenario_reference_mw, scenario_full_mw, shortfall_2030_mw, gas_nameplate_mw, gas_derate_scenario_mw
The chart’s two gas bars are not additive. 750 MW is nameplate. 525 MW is a labelled 30% event-informed derate scenario only — 750 × 0.70 — not an ELCC finding and not a claim about the proposed plant A. The cold-event record behind that haircut is in Section 6.
Against a 600 MW requirement, those are margins of:
413 − 600 = −187 · 688 − 600 = +88 · 788 − 600 = +188 MW
The planning case is reduced to 613 MW under an illustrative accreditation sensitivity, with batteries at 85% and demand response at 80%:
0.85 × 300 = 255 MW
0.80 × 150 = 120 MW
38 + 255 + 120 + 200 = 613 MW
That leaves only 13 MW of margin:
613 − 600 = 13 MW
The proposed compute sites are intentionally small, at 10–20 kW each, and are designed as heat-first nodes that can shut down during the 7–9 a.m. and 5–7 p.m. peak periods. Order 7 requires verification of these operating rules through public metering.
They contribute 0 MW to our calculations across the 413/688/788 cases.
Even if the concept eventually reaches roughly 95 MWth of recoverable reject heat, it becomes a capacity resource only after it is measured, peak-netted, and credited for its actual net electric effect.
So the argument is not that heat, batteries, demand response, interties, and compute somehow become one giant number.
It is simpler:
Heat reduces electric load.
Batteries supply electricity.
Demand response reduces electric load.
Interties supply electricity.
Compute provides additional heat and revenue.
Data sources — registry entries this figure is built from: hub_source_thermal_mwth, hub_compressor_mw, hub_delivered_thermal_exact_mwth, hub_net_electric_mw, bess_mw, dsm_mw, intertie_mw, hub_contingent_thermal_mwth, eligible_mw · computed by system_b_ledger.py
Recycled heat lowers the system load instead of producing electricity. It is counted when the heat-recovery loop is running, the buildings are connected, and the effluent stays warm during that hour. These savings go on the thermal ledger, and only the megawatts saved are credited to the electric account.
Also, the batteries hold 1,200 MWh, which is enough to cover two peak demands of 300 MW for two hours each. On the day in question, there is no midday recharging, so recharging must happen before the next severe morning event. Demand response only happens when customers answer a paid request and the load has not already been lowered. The intertie is only counted if a neighbouring power source has extra power during the same hour as our shortage.
Extreme cold can hit more than one at once, so that independence is only partial. Brandon is one plant with one failure mode. The 3-of-10 record is why that hour has to be modelled. It is not an ELCC value PD.
Data sources — registry entries this figure is built from: hub_thermal_mwth, hub_net_electric_mw, heat_pump_cop_delivered, bess_mw, dsm_mw, intertie_mw, scenario_reference_mw, shortfall_2030_mw · computed by system_b_ledger.py
Haircut the batteries to 85% and demand response to 80% and the planning case is 613 MW. That still meets the 600 MW line, but by 13 MW. Of course, an illustrative accreditation sensitivity is not an ELCC study. 85% is not dependable capacity. Hydro distinguishes dependable capacity from dependable energy. Order 1 is the real winter test, and we are asking the Board to apply it to the gas plant.
If our estimates are off and the gap is bigger than we thought, we can add more batteries or more demand response. The Brandon gas plant would be built once, at full size.
4. Winter modelling: planning stack +88 MW; illustrative accreditation +13 MW
Manitoba’s electricity use is highest in winter when it gets colder. If not managed well, adding heat pumps could make this worse. The hybrid model helps by moving some heating demand off the electric system and targeting the two daily peaks, each lasting two hours, instead of one long six-hour period.
Data sources — registry entries this figure is built from: hub_net_electric_mw, bess_mw, dsm_mw
Winter planning targets a few key hours and makes sure the battery keeps enough energy for the evening peak.
A 300 MW, 1,200 MWh battery can cover both the morning (7–9 a.m.) and evening (5–7 p.m.) peaks, each lasting two hours. Together, these times use the battery’s full 1,200 MWh capacity.
There is no need to recharge the battery during the day, but it must be recharged between successive cold days, before the next severe morning. Reservoirs supply the base energy. Hydro needs to make sure there is enough charging power and energy to refill the battery between important days. Whether this is possible in a drought year depends on timing, and is Order 1’s head-to-head modelling. An 85% haircut is not dependable capacity. Hydro distinguishes dependable capacity from dependable energy.
If the cold peak lasts more than two hours, the 300 MW battery will not be enough to meet demand during the extra time. Hydro will need to address this issue.
The sewage heat from the hub is different because it reduces the load by 38 MW all day, as long as the recovery system is working.
Claimed
- Planning stack covers 600 on nameplate arithmetic (+88 MW) only if 150 MW of unfilled demand response and 200 MW of undemonstrated winter import both show. Without the import, 488 MW does not. Conservative 413 MW does not. Illustrative accreditation +13 MW is not an ELCC study.
- Full widens the nameplate margin, +188 MW.
- Conservative shows what fails when optimism is stripped out.
Not claimed
- Summer cooling megawatts as winter firm.
- A flat 30% gas derate as proven ELCC.
- Hybrid availability above 98%.
- A 4-hour battery covering a 6-hour peak without a load-shape study.
Winter proof file
| Resource | Planning MW | The math | Still open | If it fails |
|---|---|---|---|---|
| Recycled heat | 38 | 50 MWth × coincidence − 12 MW compressor | The 37.6 MWth January extract is 4 °C × 9.4 MWth/°C from about 195 ML/d E. 9.4 is rounded (about 9.45 from the flow). Coincidence is stated at 100%. Auxiliaries are unitemized. | Without the hub, conservative is 375 MW |
| Batteries | 300 | 1,200 MWh = two 2-hour spikes | They have to put 1,200 MWh back before the next severe morning. In a 12-hour window that is 100 MW of charging power — the same battery, not a new plant. Drought-year recharge is Order 1. | We already do not claim a peak longer than two hours |
| Prospective DR | 150 | Residual after 38 MW of recycled heat; incremental to Hydro’s 310 MW DR | Inventory cells remain open. No eligible-MW, participation, or dependable-MW figures are claimed. Hydro’s 310 MW is not this 150. | Conservative already haircuts this to 75 MW |
| Intertie | 200 | 300 MW × 2/3 convention | A neighbour has to have spare power in Manitoba’s coincident winter hour. The two-thirds figure is an engineering convention, not ELCC. Order 1. | Stack is 488 MW |
| Accreditation | 613 | 38 + 255 + 120 + 200 — illustrative accreditation | Batteries at 85%, demand response at 80%, intertie left at 200. Not a portfolio ELCC. Hours in order, how the pieces fail together, duration, recharge, and whether a neighbour has power in the same hour: that is Order 1. | No substitute ELCC is published |
Those gates are the existing orders, not a second list. Order 1 runs gas, the bridge, and bridge-then-hydro through one hourly, drought-year LOLE/ELCC model, with executed import commitments folded in and imports still uncosted in this registry. Conawapa megawatts are not added to 413 / 688 / 788. Order 6 is the measured 5–10 MWth wastewater-heat pilot.
5. Money: $1.5-1.8B net of ITC vs $4.5-6.9B
Hydro’s planning assumption for Brandon gas capital is about $3.0B. That figure is still being refined. The hybrid’s overnight capital is $1.95B gross, or $1.50B if the federal Clean Technology ITC is confirmed A. We show both because credits are assumed, not fact.
The $1.95B is $0.60B batteries, $0.50B network, $0.45B hub, and $0.40B demand response E:
$0.60B + $0.50B + $0.45B + $0.40B = $1.95B E
The hub is plant and piping. The network is the district loop. Both are planning allowances for connecting buildings and for work in the municipal right-of-way. This total does not include Conawapa. It does not include buying the 200 MW winter import in the planning case. There is no firm contract in this file that would make a 200 MW winter purchase unnecessary. Until Order 1 confirms delivery, those 200 MW remain an uncosted assumption.
Overnight dollars per kilowatt are capital divided by megawatts:
$3.0B ÷ 750 MW = $4,000/kW
$1.95B ÷ 688 MW = $2,834/kW
$1.50B ÷ 688 MW = $2,180/kW
That is about 1.4 times cheaper on the gross capital.
Data sources — registry entries this figure is built from: gas_capex_b, gas_nameplate_mw, gas_derate_scenario_mw, hybrid_capex_gross_b, hybrid_capex_net_b, scenario_reference_mw
Overnight capital is not the bill. Fuel, carbon, and twenty years of running the plant are. Those all-in bands do not overlap: gas $4.5–6.9B, hybrid $1.5–1.8B net of ITC ($1.95–2.25B without). The hybrid bands are domestic capital only. They do not include a twenty-year import bill.
$4.5B − $1.8B = $2.7B
$6.9B − $1.5B = $5.4B
Under these modelled bands, the cheapest defensible gas future still costs $2.7B more than the most expensive defensible hybrid future. That gap is why the twenty-year bands decide the case, not a single dollars-per-kilowatt ratio.
Data sources — registry entries this figure is built from: gas_20yr_low_b, gas_20yr_high_b, hybrid_20yr_low_b, hybrid_20yr_high_b
No city funding agreement is named here. Who will own the pipes, and who will connect neighbourhoods, is still open. NEWPCC Project 3 has no funding agreement and no public construction schedule that extends to summer 2030.
This cost appears on bills
For ratepayers. $3–6.9 billion in fossil fuel-based capital, along with fuel and carbon costs, will appear on bills.
A smaller, fuel-free option would have a lighter impact, and a North End heat utility could bring in clean-energy investment that a standard sewage plant cannot. If heat taps are missed at NEWPCC, that opportunity is lost for a generation.
6. Audit: compare both plans, assert neither
We ask the Board to commission an independent ELCC study based on the 3-of-10 cold-event record. The historical Brandon events are evidence that tail-hour availability deserves explicit modelling; they are not themselves an ELCC calculation, and they are not a finding that gas is unreliable. Hydro's 91% start-reliability figure answers a different question: annual averages blend mild months into the average that peak planning must ignore. The Board's concern is conditional availability in the top load hours, under identical accreditation treatment for gas, storage, demand response, interties, and thermal displacement.
Data sources — registry entries this figure is built from: gas_nameplate_mw, gas_derate_scenario_mw, peak_record_mw
Against the multi-day vortex objection: reservoirs supply the days; batteries and demand response supply the hours; district heat and dual-fuel shrink the need itself. A four-technology, multi-site portfolio has no single component whose loss removes most of its capability. A single station on a single fuel line does.
7. The ask: seven orders before the concrete pours
The ask is not approval of our plan. Orders 1–4 put both plans under the same exam. Orders 5–7 keep the alternative open while that exam runs: preserve the trench, meter a heat pilot, and authorize public heat-first compute.
- Head-to-head modelling of the NEWPCC clean hybrid portfolio against the Brandon dispatchable capacity project before any capital certificate, including whether a 2030 peaker still leaves Hydro’s next northern station off the path.
- Independent cold-weather reliability audit with unit-level records at or below -30 °C and identical ELCC treatment for both plans.
- Machine-readable economic models from both parties, with named assumptions.
- Disclosure of interconnection headroom near screened anchor sites, and extreme-event performance logs.
- Preserve the NEWPCC trench option for thermal headers during the plant rebuild.
- Measured heat pilot: a 5–10 MWth effluent-stage hub at NEWPCC, with signed heat customers and public metered filings.
- Authorize public heat-first compute: locked 10–20 kW per-site band, publicly accountable, heat recovery required, public metered filings — not contingent on pre-cleared gates.
8. What would change our mind
- An ELCC study finding gas tail-hour availability above roughly 95% and battery accreditation materially below its four-hour rating would collapse the reliability differential.
- Verified capital quotes pushing the portfolio far above $1.95B gross, or gas below $3.0B, would compress the delivered-cost gap.
- A twenty-year gas scenario landing below $4.5B under filed fuel and carbon curves would break the non-overlap of the cost bands.
- District-heat uptake stalling so the hub is off would leave conservative at 375 MW, not 413. 413 still includes the 38 MW heat credit.
- If the bridge cannot cover the drought-year hour, gas stays on the table. Conawapa does not excuse that.
None of these falsifiers has been tested. Absence of a Board ELCC is not evidence the differential survives one. Each test is checkable. We have not seen symmetric candour offered for the gas plan to date.
9. Beyond winter: public compute, summer PDRC, and a ground battery
Manitoba has to cool buildings too. Cool the roofs, and store leftover heat in the ground until January.
Public compute, PDRC films, and a ground battery complement the hybrid. They are not contingent on it. The winter stack still has to stand on sewage heat, batteries, demand response, and the ties. These ideas add the rest of the year: a heat customer that can shed, roofs that dump heat to the sky, and a ground battery that banks cool until January. Together they are a year-round way to run the system so the ties point the right way. Hydro stays south on a coincident freeze. Surplus stays on the line when the Midwest is hot. That is more favourable operating conditions on import and export, not extra winter firm. They do not add a watt to 413 / 688 / 788. They do not close the 13 MW illustrative accreditation margin. Each stays on its own book until it is metered.
The third book: public compute
Public compute is a theoretical add-on. It could aid heat, and it could aid revenue. Small 10–20 kW heat-first nodes would buy heat into the loop and pay for power on hours that are not the peak. Order 7 asks the Board to let a public trial prove three rules: they shed at peak, heat recovery comes first, and the meter is public. Heat recovery at the effluent hub does not depend on them. They still add 0 MW to the winter firm ledger C.
If that model grows, the same loop could reject up to 95 MWth. That is later-scale thermal upside, not a seventh block in 688 MW. Small nodes do not produce 95 MWth. Until it is metered, peak-netted, and counted net of its own IT load, it stays off 413 / 688 / 788.
This would be public compute, not large private facilities that treat clean power as a discount. Public nodes would be a year-round heat customer, and a way to host compute that does not eat the 600 MW shortfall.
A node like that is a flexible customer and a source of heat, not a constant winter load. MISO pays for flexibility, winter capability, and clean energy. Covering Manitoba’s own heat spikes with recovered sewage heat and storage is how those exports stay real on the mornings the Midwest is also cold. Depending only on surplus sales and a new gas plant makes Manitoba a weaker, dirtier neighbour.
Data sources — registry entries this figure is built from: hub_contingent_thermal_mwth
Summer book: PDRC and a ground battery
In Manitoba, heating is the biggest issue during winter, and cooling is the main challenge in summer. Solving both problems needs a mix of solutions, not just one technology.
Passive daytime radiative cooling (PDRC) films placed on large commercial and institutional roofs could help release heat into the sky. For instance, 2.4 million square metres of roof area producing 75 W per square metre can remove about 180 MW of heat. That is 240 hectares of roof — a little under a square mile. With a chiller COP of 3.5, this saves around 50 MW of electricity that would otherwise be used for air conditioning at +35 °C E. The needed roof space is already available. These films work best during peak sunlight and do not help during cold mornings at −35 °C.
A ground battery, also called borehole thermal energy storage (BTES), stores heat underground. In summer, it lowers the chiller load by about 70 MW E. In winter, it provides stored heat to buildings, on a planning tranche of 120 MW of coincident discharge from 2035 E — later than the 2030 hour this filing is about. Right now, no capacity is available on the winter firm ledger, similar to public compute: 0 MW C. These numbers are not part of 413, 688, or 788.
In summer, the combined 50 MW and 70 MW give 120 MW of relief during heatwaves. That is set against an illustrative heatwave peak band of 3,800 to 4,200 MW; the +35 °C anchor behind the chart below is 3,514 MW E. The winter tranche above is also 120 MW, but the two are separate calculations that happen to land on the same figure, and they should not be added together. This is a separate seasonal benefit and does not solve the 600 MW winter shortfall. These steps do not help with winter power supply. They are not winter firm. They do not change the 413, 688, or 788 MW numbers.
Data sources — registry entries this figure is built from: pdrc_summer_shave_mw, btes_summer_chiller_shave_mw, btes_winter_planning_mw, btes_winter_eligible_mw, summer_pdrc_onset_cCurve shape from winter_load_base_mw, winter_heat_threshold_c, winter_heat_beta_mw_per_c.
Data sources — registry entries this figure is built from: pdrc_summer_flux_mwth, pdrc_summer_shave_mw, btes_summer_chiller_shave_mwCurve shape from winter_load_base_mw, winter_heat_threshold_c, winter_heat_beta_mw_per_c.
Data sources — registry entries this figure is built from: hub_thermal_mwth, hub_net_electric_mw, pdrc_summer_flux_mwth, pdrc_summer_shave_mw · computed by system_b_ledger.py
10. What this system actually does
Manitoba Hydro does not need more electricity in every hour. It needs more capacity during a relatively small number of difficult hours—especially on very cold winter mornings.
Manitoba is a winter-peaking system. Hydro says electricity use can be as much as 65 per cent higher in winter than in summer O. Historically, many of Hydro’s export customers have peaked in summer, when air-conditioning demand is high. That seasonal difference creates opportunities for Manitoba to sell surplus power south.
But we should not overstate that advantage. MISO is also becoming more winter-peaking as electric heating and other loads grow. On some future winter mornings, Manitoba and its neighbours may both be under pressure.
That is where the hybrid system helps.
It does not create new hydroelectric generation. It reduces and shifts Winnipeg’s demand so that Hydro has more room to operate. Depending on water conditions, transmission limits, contracts and market prices, Hydro could use that room to preserve reservoir water, avoid imports, reduce gas generation, maintain reserves or export more power. The system creates flexibility; it does not guarantee an export. Manitoba hydro is about 1.3 g CO2e/kWh, typical US generation about 350 O, so we need Manitoba not to recall its hydro from MISO at 7 a.m. That is an operating case in other hours. It is not the 200 MW import credit in the planning stack, and the two cannot be counted in the same hour.
In winter, the wastewater-heat network would supply buildings with heat while using much less electricity than resistance heating. If the proposed hub produces a verified net reduction of 38 MW during Hydro’s winter peak, it gives the system 38 MW of peak relief C. That figure must subtract the electricity used by heat pumps, pumps and controls, and it counts only while the network and its connected customers are available.
Batteries handle shorter peaks. A battery rated at 300 MW for two hours can deliver 600 MWh in one event. Covering two such events without recharging would require 1,200 MWh. It cannot be counted as 300 MW available all day. Demand response can provide additional relief, but only for loads that are enrolled, tested and actually respond when called.
In summer, radiative roofs and thermal storage reduce air-conditioning demand. The current model assigns approximately 50 MW to roof cooling and 70 MW to the ground-storage system on a typical July afternoon E. Those are summer operating assumptions—not proven heatwave capacity, and not winter firm supply.
A control platform coordinates the pieces. Using weather forecasts, building demand, storage levels, electricity prices and system constraints, it can preheat or pre-cool buildings, charge storage during lower-value hours, discharge it during peaks and briefly defer loads that can wait. Automation does not create energy. It makes existing energy and storage available at more useful times.
The emissions case also needs careful accounting. Manitoba electricity is exceptionally low-carbon, while generation displaced elsewhere may be considerably more emissions-intensive. But annual grid averages cannot tell us which generator responds to one additional export. That must be calculated hour by hour using Hydro’s dispatch model and MISO’s marginal-emissions data.
A gas plant serves a different purpose. It can produce dispatchable electricity during a prolonged winter shortage. The hybrid system may reduce the required plant size or how often it operates, but it cannot replace firm generation unless Hydro proves that its demand reductions are dependable during extreme cold, drought and simultaneous regional peaks.
The chart below should therefore be read as an illustrative load-shaping scenario—not an export forecast, a dependable-capacity study or proof of additional generation. The 38 MW winter hub, two-hour battery events, 50 MW of roof cooling and 70 MW of thermal storage remain separate operating assumptions. They do not change the 413, 688 or 788 MW supply totals used elsewhere in the plan.
Before any of these reductions are counted, Hydro must test them against actual hourly load and weather records, extreme winter and summer conditions, drought-year water availability, battery recharge requirements, customer response, rebound demand, transmission constraints, market prices and MISO’s increasingly important winter peak.
Data sources — registry entries this figure is built from: winnipeg_normals_mean_c, winnipeg_normals_min_c, winnipeg_normals_max_c, year_load_unmanaged_mw, year_load_managed_mw, hub_net_electric_mw, bess_mw, pdrc_summer_shave_mw, btes_summer_chiller_shave_mwTemps from ECCC 1981–2010 Winnipeg Richardson Int’l A. Load from the same winter and summer P(T) models as Figures 1 and 12. Battery shave is a two-hour peak-hour tool, not monthly energy. Neither series enters 413 / 688 / 788.
11. Soil & water: honest phosphorus arithmetic
The remediation math is separate to our plan, but worth consideration.
The environmental analysis stands on a binational Red River basin evidence base: 17,629 Canadian and 137,540 American water-quality observations at 3,574 stations, 121 active U.S. streamflow gauges, and 902 paired phosphorus-and-flow load calculations at the Emerson border gauge O C. The border objective is 1,400 t P/yr at Emerson O.
NEWPCC solids recovery represents roughly 4% of the required phosphorus reduction. Surplus-funded remediation bands address agricultural sources, and a full stack of measures would close the external load gap. No single measure—including the hub's heat recovery—solves the phosphorus problem alone. The arithmetic requires honest accounting of what each intervention delivers.
Data sources — registry entries this figure is built from: p_gap_t, p_border_target_t, p_newpcc_recovery_t, p_stack_midcase_t, p_surplus_range_t, remediation_floor_pct
Because no single measure closes the gap, this filing attaches a binding surplus rule the Board can require: at least 35% of heat-recovery and public-compute surplus flows to Lake Winnipeg phosphorus remediation as a first lien, changeable only by a 75% supermajority governance amendment with public notice — intergenerational equity as a lock E.
This is a water-quality case separate from the winter firm capacity argument. The energy hub proceeds on its own merits. The remediation work requires its own evidence base and its own ledger.
12. Take action
Download the filing, share the case, write a two-minute letter, or take a formal seat at the hearing.
-
STEP 1 (30 SECONDS)Read the filing
The whole technical case in one PDF, with the math addendum for Board staff. Every number in it comes from one published registry. Start with the one-page brief; then open the master if you need the full walk.
-
STEP 2 (5 MINUTES)
Edit it
This is a draft, and it is meant to be argued with in public. Every section of the filing has a thread in the forum. Quote the passage you would change, say what it should say instead, and give the reason — that is a proposal, and it is tracked from the moment you post it.
Nothing changes because it is popular. A proposal moves the filing only if it survives the same automated checks every number already passes: the registry stays consistent, the two ledgers stay separate, and the arithmetic still holds. If it does, the change lands with your name on it.
How a change lands
- Quote it. Name the document, paste the words as they stand in a
currentblock, and your version in asuggestionblock. - Propose it. Say what they should say, and why the change is right.
- Check it. The proposal is run against the filing’s tests.
- Land it. What passes is merged and the document is rebuilt.
Edit the filing in the forumGoes straight to the Drafting category.
-
STEP 3 (30 SECONDS)
Share it
Send it to someone who should see it: text, WhatsApp, email, or your socials. Share the case — the argument and the filing — or the page, if you would rather they read it here first.
-
STEP 4 (2 MINUTES)
Write a letter
One Manitoban's written comment enters the official record with no coordination: the Public Utilities Board, your MLA, or your councillor.
Take Action -
STEP 5
Stand as an intervener
For community groups and formal participants: apply under the Board's Rules of Practice and use the master filing as an evidence pack. Or bring a challenge to the open forum — every number in the filing has a category, and the falsifiers have their own.
Learn More -
STEP 6 (ONGOING)
Join the discussion
Argue with the arithmetic in public. Every part of the filing has its own category — the dual ledger, winter capacity, reliability, the money, the phosphorus math — plus a Change our mind category for evidence that breaks a claim. Bring a source and it gets answered on the record.
Qualified to check one of our engineering estimates? The peer review page lists all 48, who can settle each, and the public record of what has been found.
Join the discussion
13. Read further
The questions people actually ask, and the papers behind single claims.
Common questions
The ones people actually ask, including the objections. All questions and answers →
Is this the same Spirited Energy as Manitoba's 2006 slogan?
No. This site is Spirited Energy 2.0, an independent 2026 public filing to the Manitoba Public Utilities Board on winter capacity and recycled heat. It is not Manitoba's 2006 Spirited Energy tourism brand. The operator is Hersh Seth.
Is Manitoba Hydro building a new gas plant in Brandon?
It has applied to. The Brandon dispatchable capacity project would add about 750 MW of natural-gas combustion turbines at the existing Brandon generating station for roughly $3 billion, targeted for service in summer 2030. It is before the Public Utilities Board now, which is why the arithmetic is worth arguing about while it can still change.
Why does Manitoba need more power when it has so much hydro?
Because the problem is power, not energy. Manitoba has substantial dependable energy stored in its reservoirs; the planning problem is the additional coincident megawatts required during winter peaks. Weeks of stored energy are not interchangeable with extra megawatts at the coincident peak. What the system lacks is a few hundred megawatts during the coldest mornings and evenings. Manitoba Hydro projects a shortfall of about 600 MW by 2030 in a drought year. That is a peak-hours problem, and peak-hours problems have cheaper answers than a new power station. Conawapa-class hydro is the right later machine for that capacity path and the wrong clock for 2030. It is not on the winter stack. Gas is not needed if Order 1 shows the hour without new turbines.
What the project actually is · Where the decision stands · Section 1 — the cold morning test
Why not just build Conawapa?
Conawapa-class hydro is the right machine for extended hydro capacity, but it's on the wrong timeline for 2030. It does not fill the 2030 shortfall, it is not on the winter stack, and it is not inside the $1.95 billion. Our approach holds the coincident hour with heat, batteries, demand response, and a tested import, then lets Hydro put the next northern station back on the capacity path. Gas is not needed if Order 1 shows that hour without new turbines.
Isn't gas the only thing that works during a multi-day cold snap?
That is the strongest argument for the gas plant, and it deserves a direct answer. Reservoirs already supply the days; batteries and demand response supply the hours; recovered heat shrinks the need itself. The gas case also has to answer its own record: under questioning at the Public Utilities Board, Manitoba Hydro agreed it had derated the Brandon turbines during three of the top ten multi-day winter peak events of the past decade.
Do gas turbines work in a Manitoba cold snap? · Section 6 — compare both plans, assert neither · Challenge the reliability numbers
Doesn't sewage heat recovery stop working when it gets cold?
No — that is the point of using it. Treated effluent leaves a wastewater plant warm in every season, including January, because it is warmed by the city itself rather than by the weather. Heat pumps lift that low-grade heat into a district loop. The colder it gets outside, the more valuable the heat already in the pipe becomes.
Section 3 — how the plan works · Challenge the heat-recovery numbers
Has sewage heat recovery been done anywhere else?
Yes. Vancouver's False Creek Neighbourhood Energy Utility has heated buildings from sewage heat for more than a decade, and similar systems run in Scandinavia and at Markham in Ontario. What is unusual in Winnipeg is not the technology but the timing: the North End plant is being rebuilt right now, which is the one moment the heat taps are cheap to install.
If the turbines would only run 5% of the time, why does it matter?
Because that is the argument for testing them, not against it. Manitoba Hydro describes the plant as drought protection and capacity insurance, expected to run under 5% of an average year. Insurance is judged on whether it pays out when claimed — and under questioning at the Public Utilities Board, Hydro agreed the existing Brandon turbines were derated during three of the top ten multi-day winter peak events of the past decade. Low utilisation also makes each delivered megawatt-hour expensive: $3 billion of capital sits idle 95% of the year while ratepayers carry it. The question is not whether Manitoba needs insurance. It is whether this is the cheapest insurance that pays out on the coldest morning.
How the hearing works, and how to be heard · File as an organisation · Section 5 — twenty-year cost, not $/firm kW
Supporting whitepapers
Same registry as the filing. If a figure differs, the filing is the claim.
Companion papers — construction and the heat-host system.
- Trenchless and HDD piping analysis — share the open NEWPCC trench; drill under streets instead of opening them. That is how the district loop stays quick and bounded.
- The heat host plan — the hub, the district loop, and public compute. Heat and compute are authorized on separate tracks.
Working papers — not regulator-facing. They put the working-out on the record; the filing governs if a figure differs.
- Real total cost, line by line — the twenty-year ledger behind $4.5-6.9B for gas and $1.5-1.8B for the hybrid: operations, fuel, carbon, battery replacement.
- Five key indicators — five side-by-side measures, each with its figure and source.
- The plan at its strongest — the hardest objections we could write as an engineer, a lawyer, and a treasury analyst, and what still stands. They are ours, not Hydro’s, which is why the paper asks for a real audit.